Nick Baltas on Trend Following in 2026: Signals, Structure & Strategy | Systematic Investor | Ep.282
Summary
Trend Following: The discussion centers on rules-based trend strategies, noting a strong start to 2026 and continued strength from equities and precious metals.
Managed Futures: Significant dispersion among CTAs was driven by speed and universe choices; slower speeds and smaller universes outperformed recently but may reverse with regime shifts.
Risk Dynamics: V-shape reversals and interest-rate whipsaws were key sources of divergence; broad diversification and dynamic allocation are emphasized for resilience.
Non-Trend Signals: Mean reversion and micro indicators helped in 2025, supporting investor behavior by stabilizing allocations to trend during difficult periods.
Research Insights: New work on nonlinear time-series momentum (neural networks) and a theory paper on trend’s crisis behavior reinforce tail nonlinearity and defensive convexity.
Market Approach: The guests stress avoiding prediction and relying on disciplined, price-driven processes over narrative-based forecasts.
Access Vehicles: Consultants highlight broader access to trend via CTAs, mutual funds, UCITS, ETFs, and QIS, making implementation more flexible.
Narrative Data: There is exploratory interest in narrative/thematic data as potential early signals, with caution about conditional market impact and noise.
Transcript
Imagine spending an hour with the world's greatest traders. Imagine learning from their experiences, their successes, and their failures. Imagine no more. Welcome to Top Traders Unplugged, the place where you can learn from the best hedge fund managers in the world, so you can take your manager due diligence or investment career to the next level. Before we begin today's conversation, remember to keep two things in mind. All the discussion we will have about investment performance is about the past and past performance does not guarantee or even infer anything about future performance. Also understand that there's a significant risk of financial loss with all investment strategies and you need to request and understand the specific risks from the investment manager about their product before you make investment decisions. Here's your host, veteran hedge fund manager Neil's Krup Larson. Welcome and welcome back to this week's edition of the systematic investor series with Nick Bolters and I Neils Castro Blaston where each week we take the pulse of the global markets through the lens of a rules-based investor. And let me also say a really warm welcome. If today is your first time you're joining us and if someone who cares about you and your portfolio recommended that you tune in to the podcast, I would like to say a big thank you for sharing this episode with your friends and colleagues. It really means a lot to us. Nick, it is wonderful to be back with you uh this week. I know it's been a hectic week. Um and it's also been a little while since you and I did one of these weekly conversations, but uh how has 2026 been treating you so far? >> So far, so so far so good, I would say. Um you know, good morning, Neils. Um I would even say still happy new year. You know, in Greece, we tend to kind of extend the wishes until until math end. Sure. >> Um it's been um >> it's been a hectic start of the year, but not necessarily for just for business reasons. Um you know, we were in Greece with family and uh I'm not sure if you saw like this kind of airspace um issue that came about um in Athens. Um long story short, there was no possibility that no flights could land or take off from the Athens airport. It was a Sunday on the 4th and we're supposed to be flying back and then flight got cancelled. you know, the entire family took us 4 days to come back at the end of the day because, you know, you had the big fat Greek return. So, all the Greeks are coming back to London uh post Christmas. So, there's no direct flight at some point or at least with availability. Uh so, that was number one. And then, you know, a bit of like um uh stomach bags and stuff, you know, have have let us down specific like know the kids. So, it's been quite of a of a rough start, I would say, but you know, we keep our smiles and uh the year has started well otherwise and you know, business goes well and um we're busy, but yeah, glad to be back. Glad to be back. It's good to have you back. When you when I know you have young kids and I when you when you mention that, it kind of reminds me of of one of uh my own travels back from from uh Christmas and New Year's um many years ago when my kids were young um and where the snow uh essentially uh while we were sitting in the plane just made it impossible to get out of Copenhagen and uh and and we end up also spending a couple of days extra to before we could get tickets. But uh there we are. there are certain things we don't uh control. >> Um, nevertheless, as we always do, it's always interesting for me to besides all the topics um that we are going to talk about and they're going to be they're super great, but um I'm always curious as been sort of what's been top of mind with uh with uh with you um since we last spoke or in the last couple of weeks. So many things are going on of course, but there anything that kind of stood out for you? am a bit uh interested if you like and and and I'm spending a bit more time these days um looking into the the way that if you like investor narrative and thematics um are impacting uh as allocation or even more so um are determining um investors uh views and and you know we discuss in this um in this podcast series for years uh momentum and trend following right and and what is momentum trend and following at the end of the day it's just perhaps um an expression of uh of slow digestion of information. >> Mhm. >> And you know we see that in the price path which allows to then document a positive or a negative move that in itself supposing that we have this kind of continued digestion of information at a slow pace continues and that's for us a tri a signal. So what happens before we kind of digest this information at a slow pace? There is clearly information that is produced and and news that is produced in the market and and and information that gets disseminated via the media and the you know the the internet. Um so that cloud of information if we end up achieving to get hold of prior to the price formation happening it's almost as if we'll get the momentum before the momentum itself plays out in the prices. So, long story short, um I'm actually quite engaged and and and focused on um um on a partnership we're doing with a kind of a a vendor if you like or or an aggregator of um of this type of data of narrative data. Um and and I'm I'm almost trying to think how this information or this type of data set can be utilized for signaling for aocation purposes, for strategy design, you know, you name it. But you know if you ask me then you know what what what keeps me excited or maybe what um keeps me thinking um in a in a productive manner that's certainly one um not a concrete answer but certainly a concrete framework or or maybe idea um that I'm I'm kind of spending my days when I'm kind of running these days or I'm doing things outside of just replying to emails or or or pitching to clients. There you are. That's the one thing. >> Yeah. No, it's kind of interesting. it it is slightly related to something that I had written down um not like something that's uh you know top of mind but still interesting. So for example, if I had if if let's just say that the narrative um had been oh Japan has increased uh its interest rates uh by a lot clearly that must mean that the yen goes up right let's just say that for an example and um I just saw a chart uh this morning um that that showed uh how much actually of an increase the the Japanese have been um seeing in the interest rates. At the same time, the yen has just done nothing in terms of strengthening. In fact, it's probably as low as it's been, more or less. So, so I I understand the interest in this thing about what if we could get hold of the momentum before it happens, so to speak. Um but I'm also reminded from last year's um price action um in a world that is getting crazier by the day how important it is not to have any kind of prediction in your in your model. Um, so I'm kind of doubling down on this um, trend following where a lot of people don't like the fact that we have no story and we have no um, view as such. Um, but just the fact that we're super disciplined at following what has happened rather than trying to get ahead of ourselves. I I probably reverse the argument a bit in the sense that I I hear you on the ex on the example about about Japan, right? And I guess after the fact u maybe the I guess the explanation there is no exposure or maybe there is no sensitivity >> to the narrative itself. And and I think that's maybe the direction I'm I'm thinking of going in in terms of there could be some story around whatever let's not name it but let's say an event the question is you know whether there is at the minimum any price reaction that we should pay attention to or not. So you know the fact that there is more intensity on a specific topic does not necessarily translate to the fact that you know price response will be significant. So I I don't have an answer, right? I'm no >> thinking out loud, right? Let's say for instance um you know equity prices are not necessarily impacted too much about inflation unless inflation is rising beyond a specific threshold. >> So it's a bit more of a conditional exposure and that exposure to the extent that we can harvest it and and and monitor it and quantify it then we might have a chance of making use of it. Um I'm not necessarily sure that there is an end to it but I'm actually kind of experimenting. The other way that I would think about it to your point and I think we've discussed it quite a few times because at the minimum I'm I'm I'm quite of a proponent of um of this idea that signal to noise quantification is important specifically for the Vshapes. So how do we how do we how do we get hold of signal to noise ratio in the market is not too easy right I mean beyond the returns scaled by volatilities which is a return over volatility or like signal to noise if you think about it in a more kind of um huristic manner we have to get hold of um um I guess non-pric information and and maybe recycling of of themes and and and narratives could allow us to get there. I don't have an answer by all means. Uh but that's a place I'm I'm quite excited about. Um just because I think we've done as much as we could with prices and volatilities and maybe there is more to be done outside of that spectrum. Maybe there isn't. But unless we experiment, you know, we we will we will not be able to to to to bring news and innovate. So that's the topic. But by all means, I think challenges are are there for us to at the minimum um um if you like entertain. So >> everything is well received. >> Speaking of narratives, you you're not I'm sure you won't believe this, but but it is it is true. Um so I was looking at my notes from last time we've recorded one of these episodes, and I don't remember exactly the date when this was, but it's been a few months ago because >> probably October maybe. >> Yeah, something like that. And so I always see what was on my radar back then. I had written down two things. One was Greenland. The other one was Trump's attack on the Fed. And I'm thinking what? I can just reuse these topics and say, "Oh, this is very much on my on my agenda." And I know you can't talk about uh politics uh of course, but um >> maybe maybe give me maybe your give your audience the signals then now like the m is m is yours. >> This is the thing. I have no idea what the narrative or what the markets would even do if we had um uh you know a massive outcome in in either of two of those two. Um but anyway and what what would actually what is interesting relating to to these two of course is that the the person at the center of attention in in both of these cases um he is coming to Switzerland to the to Davos as far as I understand. So, so, uh, there'll be lots of narratives coming out of, uh, of the of the mountainous country, uh, I'm sure in the in the next week or two. >> I'm sure. I'm sure >> people will pay attention for sure. >> Anyways, let's move on to something where we are uh, where we can talk freely, namely, namely trend following. Um, so January is actually off to a very strong start and, um, I hate to say it, so was January of 2025. Um but there should be no necessarily repeat just because we uh see both starts being very strong. Now January 2026 though nevertheless it is kind of the seventh month in a row where we're seeing very strong conditions uh for trend. We're also seeing a continuation, I think it's fair to say, of what really drove performance in the second half, namely medals uh and equities. >> Certainly setting the pace, but obviously I'd love to hear your uh kind of takeaway from um the the beginnings of 26 and and maybe you want to talk a little bit about um you know, 2025 and and then maybe later we can talk a little bit more about the QIS side of things. But just if we just stick to trend right now um what what you've experienced. >> If we stick to trend um I think you probably said it all in the sense that I think 2025 I don't have the stats but this is more of a hunch um is possibly one of the largest dispersion years we've seen. >> Mhm. >> Um without necessarily that meaning that we have seen >> um correlation breakdown between th programs. So specifically if we you know if we look into more of an outlier type of a of a scaled correlation analysis in other words to make it very very simple if a drop came in April I'm sure it came across the board maybe a different scale uh but if we assume that the scale is somehow um diluted in a correlation analysis and there are ways that we can do so throughout the year whether you're following like you know a 3 month or a six month or a 12 month trend like you know shortterm to medium to a longer term the correlations were relatively elevated anyway because to your point there were a few markets that exhibited strong trends and this is probably equities and then precious metals that's it >> y >> um and then there were pretty much the same markets that actually failed and maybe it's now um quite continued failure uh for some time you know interest rates obviously that's the one I'm talking about that's the elephant in the room >> um >> so even If the moves were captured relatively successfully if you like by a variety of trend speeds and programs the dispersion of outcomes was quite significant. I guess from that perspective, do we learn anything? You know, frankly, nothing more than just yet another year that we know we do the diagnostics and and we realize that, you know, yes, if you were too fast and possibly that Vshape did not play at your benefit or if you were too diversified, you're actually missed those few markets that did the job and the rest, you know, we're just now adding noise to your program and diversification did not play out for this year and maybe that's what we have seen the last couple of years. Are we entering a phase of should we think differently now from this point onwards? you know, the fact that we discussed that at the year end um at the year end um >> conversations, yeah, >> conversation we had with with the rest of the crew. >> Um I think the universe and the speed are probably the two most important um pillars of dispersion at least for the last year. Um, and it's fair to say that had you been more on the faster front and had you been more diversified or if you like a broader universe, probably you wouldn't be at the top ranks. But the fact that we now have a universe of very positive and very negative returns that in itself is is is an interesting observation as to how we react as product developers, how do we react as consultants, how do we react as investors, how do we react as, you know, asset owners. Let me make a pause here. >> Yeah. >> Um I can I can move on more to what we're seeing and and how we're discussing with specific investors, but maybe if you ask a question, I just put your views. I can I can I can pass it on. >> Let's do that in a second because let me it gives me a chance to do two things. One, just to give a quick uh trend update as I normally do on that, but also just to reiterate something I said to Rich last week. Um >> and by the way, I think Rich brought a really really interesting um set of topics uh last week. So for those who missed that conversation uh you know you should really go back and and and listen to that. But anyways, um, in so many ways you you talked about the dispersion, all of that, but in so many ways you could say that trend, uh, sorry, the last year was, you know, unusual, extraordinary, unpredictable, all of those acronyms that we we we we hear. But in a but in another way, the year was also very very familiar from a trend following perspective. Now I know the outcome there were differences but if you think about the fact that most managers probably made all the money in only a handful of markets and other than that we had many small losses maybe one or two bigger losses in in markets. I mean that's very classical trend following. Um so that that is to me is very comforting. um despite you know whether people ended up plus 5, plus 10 or minus 5 doesn't really matter but but the structure of how the year ended and and the return distribution all of that stuff actually look pretty pretty uh familiar to uh to um to trend followers. Anyways, we talked about uh trend uh sorry CTA coming off to a good start so far this year. So let me quickly run through the numbers. We have the Btop 50 index uh as of Wednesday evening and by the way I think yesterday was a pretty um quiet day but as of yesterday uh Wednesday evening up 374% for the Btop 50. Um so gen CTA index up uh 4 and a4% very strong trend index even more so up 4.67% and the short-term traders index up 1.75%. So all coming off uh very strong. Now, I actually completely forgot to look at the traditional markets, equity markets, and and so on and so forth. Um, but so far, just eyeballing it a little bit, obviously, we've had some positive developments in in many of the uh equity markets already this year. Um, while fixed income is probably um only slightly up, uh if anything, so we'll we'll leave it at that. Now, before we get into the topics that we're going to be talking about, uh Nick, um just uh allow me to uh mention to uh our listeners that I just published the eighth edition of uh my ultimate guide to the best investment books of all times. Um and that essentially now has more than 600 book title titles that people can uh dive into. So, it's more of a reference guide if they want some inspiration. Uh hopefully now that we put it in in one um PDF, it makes it easier for people to grab hold of. If you don't if you're not on any of my email lists, um you can go to toptradersunplugged.com/ultimate and you can get your copy and it is of course free of charge. All right. Now, I would love to um hear you talk about 2025 maybe from a QIS uh perspective. Let's break it down maybe into sort of universe versus uh speed. A little bit of continuation of the conversation we had uh with the uh the rest of the crew as you said um back in December. >> Let's do that. And and by the way, as you were speaking, I was kind of checking some of my stats for for the month to date or like year to date. Mhm. >> Um >> yeah, strong performance broadly speaking like on the on the four to five to 6% depending on the speed. Um and that's more for a 15 volt type of um of a measurement primarily driven by equities and then commod. So kind of similar things to what we saw in the kind of second half of of of the year. Let me pull out some details for for the discussion, right? Um so you asked about um I guess speed and universe. Speed is probably the most important differentiator and then universe is the second. >> If we kind of stick to the core design principles of trend following. Yes, whether it's dynamic sizing or or static is important. Yes, whether we use correlations or not is important. But long story short, assuming that there is some form of stability of volatilities and correlations, I think the speed will create more of a dispersion than the sizing. I haven't tested it but that's my hunt as it stands. Um, so we did this nice exercise which I very softly discussed in the in the in the group review. Um, that looks into fast programs. Let's say looking into like a twomonth type of a signal you know on a on a half-life basis that goes up to 12. Um, and that was kind of similar to what Quantica did, which I think also we kind of briefly touched upon back in December. Um, and then we did a similar exercise whereby we fixed the speed, uh, but then we changed the universe and we went from something very concentrated. >> Um, and I believe we use something similar to DBMF's universe just to have like, you know, an anchor which is kind of independent to our choices um, up to something that is very broad, you know, more than 80 100 markets or so. And then we tried to have some sort of mediumcaled universes maybe a bit more liquid but actually well thought through like in the sense of know if you use for instance non- US equity indices you might as well just use the EF or you can use you know UK Germany I don't know you name it like four or five major markets XUS so you can broaden up a small universe by just having maybe more tenors in specific interest rate markets or you have more in that sense. So the if you like the medium-sized universes would be reasonable rather than saying oh let me get you know 20 currencies and three equities and and we try to see here and here from for the last 25 years if you were to rank purely by annual performance however big or small the differences are which was the universe that outperformed and which was the speed that outperformed and then do some sort of a ranking which was the best speed the second best speed the third best speed and so on and so forth. there's something quite striking coming out frankly um and this is that for the last three years and that's exactly what I said in the in the December discussion if you were slow you would have out outperformed >> there is no other time in those 25 years that three years in a row being so slow would have allowed you to outperform and by the way the second best speed is just 9 months instead of like 12 months So slow has been the thing >> m >> for SVB for dollar yen for liberation day the three of these shapes you know you and I have been discussing for the last I don't know how however many months now universe wise having been small is possibly quite dominantly the outperformer for the last four years um if it's not the first to second best which is also the first time we're actually seeing it historically. There have been other times that being small was actually helpful like I think for example 2018 which I would imagine would have helped you uh towards Q4 of 2018. Um but there have been years in between whereby being very small and concentrated would have delivered the worst performance. It's only the last three four years that a concentrated universe is top ranked so stably um visav history. Right. So we have two if you like uh stylized facts that we have not seen historically and they're almost the flip side of what we saw maybe in you know in 2021 or so. >> Can I be even more explicit because you shared you shared the data with me. What's what's what really is striking to me when you talk about the portfolio size because I think and again I'm not trying to pick a fight here with my good friends in the replicator space but clearly replicators tend to use smaller universes even though that we have seen uh some some uh some new uh entrances using larger universes. But what is very interesting is as you rightly say that pretty much since 2016 small universes have been um you know the better performer. It's flipped between small and and super and super large meaning 18 markets or so. Right. Okay. But but >> dominantly it's been it's been that. >> But then what was really striking is that if you go back to the environment uh post the tech bubble. >> Yeah. Exactly. and all the way through 2009 >> exactly >> being small or trading a small universe was the worst universe every single year for something like seven years in a row. So what I'm or what I'm trying to say here is that I think we need to be very cautious of making too many conclusions based on people saying oh but look we can and here I may be rep maybe picking a little bit of a fight with replicators by by replicators coming out saying oh look at this we've now demonstrated for the last three or four or five years that we can outperform uh the CTA index. Sure, that is true and it is in line with what you're showing in terms of your analysis, right? But if we were to expand this over a 25 year period, I'm not so sure that that would have been the same conclusion based on your completely independent type of data here. >> Yeah. Yeah. I think that translates into a question which is do we learn anything uh on the back end of of this analysis? You know, in all fairness, know my mindset has always been try to be reactive, try to be diversified. >> Yeah. Did we get it wrong? I would not claim that being a wrong decision, but certainly from a performance standpoint, this is not necessarily the one that favored us, you know, made us a favor. Yes, it did in 2022. Uh at the minimum from from a reactivity standpoint, >> but um >> but that's one if I can interrupt here, that's the other thing that I find interesting about um your data when we look at the the speed, right? >> Yeah. So in 2022 you you conclude based on your analysis that actually being very were very quick uh was the best. Um and of course you can't you not you know you can't see whether it's just a you know how much of the difference is you can just see you know you can't but exactly but interestingly enough 2022 from from our side at done uh was actually one of the strongest years we had seen in in a long time. So, so that that also uh strike me as being very interesting uh as well. Um and I don't know what the short-term traders index did in 2022, but also in 2008, I do remember from our analysis that yes, the the very best single look back period was something like 56 days or whatever it was. Not a very particular long period, but every single period did well that year. I mean, you know, in a >> So, that was my point, right? That was my point. My point here is that >> so two things given we had sustained trends in 2022. >> Mhm. >> So we had the commod in the beginning rates throughout and then a dollar in the summer until November. I mean you get to it a month later or a month earlier fine. Uh it will create differences but then you know from the plus 40 to the you know plus 35 you know fine you can still rank them but in reality in terms of absolute returns that's actually quite substantial. So what this analysis is missing at this point in time is also an indication of what's maybe the spread between the top and the bottom ranked >> that I might end up kind of adding right. So to give an indication of that what is the spread in that ranking because it it might actually be like very very small for that to matter. >> Mhm. >> Um but you know setting that aside I think I I think it's a useful >> kind of um grid for us to >> to get a better sense of of the impact of those decisions. But as I keep on saying and I've discussed that with you and I've discussed that back in the days with with Morris when we talk about how we think of of str design in the QI space I think reacting to out or underperformance is is the wrong recipe right uh either way >> um I think discipline is important I think trust to the process is important um and and if there's anything to be learned frankly um it's not that a specific universe is better than another um it is more about that the Vshapes are the ones to have caused this dispersion rather than oh actually you should have been longer term because there are longerterm trends. No, that's not what's going on here. What is actually going on is something as much more subtle and it's a reversion. It's not that the that the signals and the and and and the directional moves that we have seen had like a longer duration or a shorter duration and therefore a certain speed is better than another. there are beyond trend dynamics that we should look into rather than completely amending our if you like our philosophies about strate design and and and reacting to it because you know how it's going to be right you know imagine being here in uh I don't know in whatever 2022 and and say hey I'm going to run it you know fast because you know I think that's the best and in reality you know you just make the the wrong decision at the wrong time so it's not about reading back the past but you know I think some of those fact patterns are actually quite interesting at the minimum to to to be aware of that that was the whole point of that analysis, right? >> And and I would add to that because I did my own little different type of analysis as as I knew we're going to talk a little bit about that. I was trying I was digging into some of the data I have access just to see kind of a little bit about kind of contribution by sectors because a lot of people uh in 2020 after 2022 concluded, oh yes, we should just continue to to trade mostly fixed income markets because they have been the best ones for the last 20 years and they were fantastic in 2022. We should just continue. Well, lo and behold, um pretty much since 2022, it's been kind of the worst sector to trade. Um and and so again, it just goes to this uh idea of not trying to uh uh predict too much about what the future uh will bring. Um the next thing I did was I looked at kind of all the sectors that we trade. And it's not necessarily the same um definitions that people might have. Um but it is interesting to see uh if I go back about 20 years that all sectors have actually been profitable. If you look back uh then we've been able to find opportunities in all sectors but it's clearly changing over time. Um and and um and so the the true diversification sort of the the truly diversified portfolios um I still think uh over time is a is a better is maybe a strong word to use but it's um uh to me is a more um intuitive way of doing trend following is to have the full breath of of markets in in the portfolio and not trying to >> be too concentrated. >> Well, I'm I'm in favor of of of dynamic allocation. Um, and if the opportunity set allows you to enter into a space that at least seems to be delivering some return, you might as well just go, right? So, so I see it more as an opportunity set that, you know, to the extent it's available to us, we tilt in favor or against depending on the intertemporal trend uh trend signals. So, I'm with you on this one >> perhaps with this conditioning information. >> Sure. Cool. All right. Well, we have three papers. We'll see if we get through all of them. Um, but I think we we will certainly touch on all three. Um, they're very uh different. Um, and the first paper actually they all came out I think in December uh of uh of last year. So they're relatively new. Uh the first one is probably the most accessible for most people to read. Um it's from Mikita the consultants. Um people might remember uh a paper we discussed a while back um certainly on the podcast not exactly sure who was who I spoke to about where they talked about trend following in the light of sort of crisis alpha and um they introduced this term I think they were the ones who introduced the term about first responders second responders and so on and so forth >> the RMS framework >> yes exactly and um and actually I think that was very helpful um frankly um I think it helped people to understand uh when um they should expect um positive performance during crisis periods and and so on and so forth. So I'll let you go through um this paper. We don't want to repeat uh things we've talked about too much, but still it's always useful um especially when it comes from people who probably have a lot of their or have the ear of a lot of large institutions to to uh hear about how they talk about trend and and what they're focusing on. So I'd love for you to maybe pull out some of the highlights in in in your perspective. >> Yes. Um and and as you said, that's more of an introductory paper in in in this regard. So there is um at least for for for the for the connoisseur of the space. Um it's more of a reiteration of what following is you know what's the the reason of its existence and and how we can get hold of it and so on and so forth. Um and you know speaking obviously of the RMS framework of course very well known it's pretty much kind of in line with our broad principles you know we wrote a paper back in 2019 as to how we think of defensive solutions and what are the different pillars that allow us to um to to to deliver defensive solutions from the most reactive and high basis and very expensive uh from a cost of carry standpoint to the more carry generating but not necessarily adding to the downside and obviously trend following is a kind of intermediate component uh so what how we had done independently at the time kind of aligns quite well with uh with their framework. So effectively what they came up about I think to me it's more interesting the fact that themselves put out a paper on trend uh they generally speak about themes and bigger topics and aggregation of of um of exposures and >> doing one piece on trend it's probably an indication of um I guess an increased uh interest in the space >> uh which I think was interesting for for us to at the minimum kind of bring up um they talk about what trend following is. I'm not going to go into the details. I think there's a nice section that talks about the differences in the programs which by all means I'm just going to read them out um list them effectively because we we kind of know and we have discussed them. So universe we just talk about it. >> Speed we just talk about it. >> Diversifying non signal non-trend signals. I'm going to come back to that because that's quite interesting. Stop-loss and profit uh taking triggers. um static risk allocation or dynamic risk allocation or maybe equal risk or risk budgeting between assets or asset classes that's an important one we we have also discussed it and I think I've just mentioned it earlier on on this dynamic allocation the opportunity set and then the impact of volatility targeting leverage that's another topic I think you and I discussed a few months back as to how we feel and we see the impact to leverage coming from portfolio design and um and and and Vshape dynamics and correlation shift. So they kind of mention a few of those topics as being important for um I guess for the for the design of a strategy. >> I'll probably spend a kind of tiny bit of time also kind of related to your question on on on on QIS and the CDA space from our um vantage point in 2025 which are the non-trend signals. Um I think I've been uh a proponent of of thinking about me reversion dynamics, micro indicators, K trades and so on and so forth. Um you know 2025 was a strong strong strong year for all those concepts. Um they all performed really well. They all performed in line with um at least the premise of diversifying trend following. So that did make I think a significant uh or had a significant impact in this dispersion specifically for those that implicitly use them um in the design um and I think they can have some value um not to make trend following better but allow it to be maintained in an location at times that it's actually not performing well um and and even if it's understandable under performance to your Like yes, you're holding equities. What do you think is going to happen when equities fall by 10%. Like 1 plus 1 equals two. I'm with you on this one. >> It's more the point that ultimately investors follow like a like a a an increasing concave utility function to put the economic term to it, which basically means that you know we enjoy gains less uh than the pain that we incur when a similar magnitude of a loss um is is coming. um and withholding a loss uh is tougher. So to my earlier point, should we be able to allow allocations in trend following be maintained by those non-trend components, I think we eventually make the investment process uh more complete in in the longer term and I think that's important. Um I think the last point that I would mention about the Makita paper is that they are they're actually quite open by the fact that now institutions have a variety of possibilities in terms of accessing the the style. Um of course they talk about CTAs but then they talk about mutual funds, they talk about usage, they talk about ETFs, they even talk about QIS. Um so I think to me it's a recognition that the possibilities of accessing it have widened and and and broadened. Uh but also the need for it has become more mainstream than before. Now whether that is a policy portfolio discussion or whether that is still an overlay discussion, it's um not of secondary importance but maybe of a secondary topic that is not necessarily covered here but at the minimum being there, you know, it's an indication of of of of interest and and I think I just wanted to bring it up um in our discussion. That's what it talks about. So I think people as you said uh I'm sure they'll find it very easy to go through as a as a quick overview. >> Yeah. No, absolutely. And and uh I agree. I think that um I mean you put it in this in in the sense of of diversifying uh signals to in order to maintain an allocation to trend. Um and I think also from memory at least um it's been a little while since I I read the paper in in detail but I think it's this idea also that a lot of trend following allocation kind of fails mostly because of the behavior or mostly because of the kind of reaction pattern of of investors. Um and and therefore it is definitely a uh an important uh an important point uh for sure. Now I think we are going to um move on to the little bit more hardcore uh stuff um where you need to be the translator of these uh papers. Uh it's actually two papers we're going to talk about. We'll probably come back to them given the fact that uh we we feel that they are very important. They deserve uh respect and and maybe um we haven't been able to uh devote enough time to really get into the nitty-gritty of that. I'll I'll I'll let you give your own disclaimer since you're going to be the uh the the the narrator or the translator of these um very important papers. But the first paper is by a group of distinguished people uh I'm sure um uh to be as Moscow uh who among other things uh it seems like he's linked to uh our friends at AQR uh we have someone from the Stockholm School of Economics Ricardo uh Sabatuchi Andrea Tamonei uh University of Notradam and Bian Ulf um at University of Hamburg. So this paper also came out in December of 2025. >> Um, and it's uh title is nonlinear time series momentum. That's as much as I'm going to say about it. Then I'm going to turn it over to to the expert. Um, and since I'm not a quant, I'll be allowed to ask some stupid questions along the way, I'm sure. >> Um, no, thanks. Thanks, Neils. Um I'm I'm sure this paper has appeared on people's um kind of inboxes. Um you know it came out in a sarin just before Christmas. Um of course Toby Moskovich was the you know was one of the co-authors of maybe the more um wellsighted paper when it comes to trend following in the academic literature. That's the time series momentum from 2012 >> right >> um with his colleagues at AQR at the time or in um Peterson Lassa Peterson. >> Yeah. Um so it's called kind of time you know nonlinear time series momentum. Um it's quite catchy to talk about nonlinearities in the trend following space. Um I think my honest opinion before we go into the details is that they spend time discussing a topic um but I don't think it's necessarily completely new um and this is how past information predict future information and whether that is a linear or a nonlinear transformation and I'm going to go into details. I think the novel thing that they bring about is that you know if there are those nonlinearities instead of us forcing them by some sort of a parametric approach let's say we know we utilize something that is called a sigmoid so it's a signal that goes from negative to positive but you know it flattens out in the tails their point is that we know maybe we let the data speak and maybe we use a neural network that allows us to uncover those relationships and then we can determine the positioning on the basis of um uh of those models. So let's kind of break it down um into into I guess into pieces into components. I guess let's think of um of of trading signals in the trend space. Um what are the possibilities? I would probably say that the simplest of them all is a binary signal. >> The market goes up, you buy. The market goes down, you sell. And maybe you size your exposure statically or dynamically by some level of all. Fine. Let's assume that all volatilities are the same. So you just buy a unit of oil when oil goes up and you sell a unit of S&P if S&P falls. That's it. Assuming that they have the same wall. Right? Now maybe one step of I guess of departure from that is to say well maybe if my signal is too close to zero then yes the sign is positive but in fact I don't have so much confidence on it. So I might as well just reduce my bet as a function of how big that signal is. In other words, if it's a basis point positive, then maybe I should do a basis point allocation rather than a 1% allocation. And if it's, you know, 1% of a positive return, then maybe I should do a 1% allocation. So I should scale linearly my exposure to the signal estimation. And that moves us from a binary mindset into a linear mindset. The bigger the maggot of the signal, the bigger the risk or the sizing that I should do. Um but there comes now kind of the third iteration of of of that signaling which again has been uh studied in some form or fashion. Maybe not as much in academia even though there are some nice papers that do some uh empirical analysis on that on that topic. But certainly I would want to believe in the industry. I mean certainly if I were to speak about um the work that we do um is to acknowledge the fact that yes fine you might actually use the magnitude to scale but how about a an extremely positive or an extremely negative signal. Let's say now you witness S&P up by you know three units of sharp ratio. How confident are you to size your exposure so much that these three is going to kind of repeat? >> So there comes a point whereby confidence and maybe estimation noise and maybe reversion dynamics come into play and then you kind of say like you know what there is a point whereby I just want to flatten out my exposure. So I get it, the higher the better, but you know, probably there should be some sort of a concave form on the positivity side and more of a kind of a convex profile on the negative side that flattens out the exposure. So I still maintain my linearity around zero. So positives are positive, negative, and negatives. But then in the extremes, I should just flatten it out. And I call it a sigmoid. Um, you know, they call it the vapor sshaped. Um I'm using the kind of the Greek if you like translation of what um of what an S is called like the sigma. Um so that's if you like the third iteration. So again to repeat binary long or short linear just use the magnitude sigmoid maintain the magnitude but then bring some nonlinearity. Um and if we stick to those three before we go to what the paper really really focuses on, there is some form of um of a benefit performance-wise. If you look into empirical analysis, going into that more like of a sigmoid type of mindset or going into this kind of linear space, um maybe the only places that you'll see some loss of performance is when a tiny bit of a trend is enough for you to get full into the position with a binary signal and then benefit from an early start. >> Mhm. >> But imagine a signal that goes bit positive, bit negative, a bit positive, bit negative. with two full positive and two full negative exposure you'll end up acrewing a lot of costs. So net of costs it can be a debate how the binary signal can actually help. So this transition is always a bit helpful. Um so there is empirical evidence and they also showcase it that you know this nonlinearity even if we force it to be like a sigmoid and there are variety of ways that we can produce that mathematically and and parametrically um you get good performance or better performance. Their point is why would you have to force the shape of that sigmoid? As in everything I've discussed so far is not necessarily driven by the data. It's almost as if we're forcing it. So the sign of past return is our determination. The linear is our determination. The smoothening of the tails is our determination. And the signal we put in place is our design. And yes, maybe after the fact we look into the empirical stats and say hey you're actually doing better that in itself almost says that yes probably the data generating process is such whereby you have this continuation of return but in the tails it ends up reverting. So they say why don't we let the data speak. >> Mhm. So instead of us scaling up and then smoothing it in the tails or scaling down on the negatives and then smoothing it out in the tails, let different markets you know be it equities or rates or commodities and so on and so forth force this nonlinearity while still preserving and that's important the fact that the signal is positive I buy the signal is negative I sell. And that's important why because the minute we depart from it it is no longer just a trend strategy. Imagine me being here and telling you, "Hey, your signal is positive. You're going to short." Well, this is no longer a trend strategy. It's something of a mixture. >> So, they use a neural network. I don't want to bother you too much with the details, but they use a neural network to effectively um um if you like extract the way that past return translates into future return, right? What's the what's the transfer function? How do I um document a signal? And how do I determine my position? And should my position be linearly related to my signal? Should it be fading out if it's too extreme? Should it maybe just fall? Let's say to my earlier point, if my sharp ratio is three, should I possibly just go to zero? I think crossing zero and going to negative. I see the value of that because if the data tells you you need to actually go short then it basically tells you that there are reversion dynamics. Now whether we bring the reversion dynamics into a trend system or we use them as a separate engine to my earlier point that's maybe cleaner from an estallocation and from an from from a performance attribution standpoint but they do show I guess to to go to go probably closer to to to the end of um of my overview that um empirically the data spits out transfer functions that have the nonlinearity that we're discussing which is the smoothening in the tails but even you can end up having reversion dynamics as well in the very extremes. Um now I said that those pictures of fitting data at the end of the day just like know if you put past returns future returns and do like a nonp parametric analysis you'll see it. So there are some papers a couple of years back I believe some of them we have discussed here but maybe my memory is kind of failing me now that show this relationship but what they say is that they are now actively monitoring it and fitting it to to get to that point. So to conclude basically the paper says look there are nonlinearities those nonlinearities exist in the tails um this is something that we should at the minimum be aware of whether we design that parametrically or nonp parametrically it's a good question whether the neural network is bringing value maybe it brings value um you know that's why one of the findings that you know the the neural network agrees with the theory um and the fact that you know in the tails possibly the estimation error is creator and so on and so forth and some reversion dynamics are there Um I would probably say it is an important discussion to be had as to whether this incremental complexity will bring significant value versus having a sigmoid that in itself kind of be in itself behaves um in line with with with expectations. So I leave it uh in in this regard but by all means you know very good work and and and I'm sure people will will spend time looking at it and maybe trying it. >> Okay. Well, the last paper um is from a uh a fellow Dane actually. Um so uh um so we're going to be talking about a paper from a gentleman called Christian who works at the uh one of the largest pension funds in Denmark called ATP and he also posted a new paper in December called on the autonom autonomy of trend. So, um I know we're going to have a hard stop in our conversation today. So, I'm going to let you talk as much or as little uh about it. I think it is uh important because it touches on some of the key things that people actually uh consider uh when they incorporate trend following into their portfolio. Uh and that is kind of the again the defensive uh nature and mechanisms. So, u I'll let you uh I'll let you talk about it uh Nick and we'll see how we go. Yeah. Yeah. Yeah. So that's um that's a nice one by by Christian. So I I I I also happen to know Christian. Um and um it's a nice piece of work. It's actually quite mathematical. Uh it's more theoretical. Um but I think what is what is important at times with those theory papers is to kind of look through um how beautiful sometimes math can be and kind of identifying what we experience in reality and kind of giving a bit of a fundamental base if you like. Um so his point is more about okay what is actually driving trend following and and how do they behave across different market environments and what is the underlying process of the data that gives rise to trend following um and and and somebody could argue yet that you have autocorrelation returns you know positive is leading to positive negative negative there are some good papers from back in the days that showcase that even if returns are independent if volatility in itself is serally dependent you can still benefit from a trend following strategy uh I'm not going to go into the details of this one but you So that was the first that came to mind when I saw when I saw the paper. Um they you know he he he kind of built a model you know he says that no log returns are following a ocean process. Um you know we we we design a trend signal which is just the summation of past the returns. I'm not going to bother you too much. You know they're using a binary signal u to go to to back to to our previous conversation. But then what what he what what they try or what he tries to to come about is um a mathematical explanation of what is the expected payoff of a trend following strategy unconditionally and then importantly conditionally on what has happened in the in the recent past which is more of a of a trend following behavior in itself um specifically condition upon um some cumulative return of a specific horizon. And I think the direction of travel here is to say well if I'm observing a significant draw down what is my expected payoff from being a trend follower. Um and there are two kind of components that come out of this analysis which is one being the directional component. By all means we know what that means. You know if you get on average the direction right you will do well. So if you're holding equities more often than not with an equity risk premium that is positive over the longer term you'll probably do well. And that was I think the challenge we had with equities and bonds you know three four five years ago. I'm sure you remember the days that everyone said, "Hey, you're buying equities, you're buying bonds, you're just benefiting from carry and equities premium." No longer the case. I think that that that doubt is is is out of the window for now. Uh but then the second element which is also more important for for us is the timing element. So how quickly can you turn into a situation whereby your um price process um is aligned in terms of direction with your signaling. Um and I think one of the more interesting things that you know his analysis uh showcases is that you know beyond the dependence that we have and we know that you know the getting the signal direction right is is is the recipe for success and so on so forth. Um there's also this part of the analys that says even if returns of the process you're trying to allocate to on the base of trend following is um is what we call iid. So in other words independent identically distributed. So every single day the equity returns are drawn from the same distribution but today's value has no dependence to yesterday's value and so on and so forth. So whatever happens every single day is independent. Conditional upon observing a drawdown which can be a consequence of independent draws like you know I can do heads or tails 10 times and I can get heads 10 times. They are still independent right. But it can happen. >> Sure. Um they show that you know um beyond specific thresholds mathematically defined a trend follower can still um deliver positive return and specifically in the draw down specifically as a crisis alpha uh which goes kind of hand inhand >> um with um with some of the empirical observations. So I kind of I know I went a bit quick um in the interest of time but two things to flag. It is a mathematically heavy paper which is actually quite nice. Um but I think what's more important is to associate now the findings that um if you like formula and and variables um if you like contain to to the experience we observing and and and I think going past the fact that you know serial correlation is important and the fact that even a an IID process can allow you still to be profitable with a trend following strategy is an important finding. Again, it reminds me of that work that says conditional dependence in the volatility space is enough for you to benefit from a trend following strategy or in other words to predict the directionality. Um so I don't know if you have any reflections upon that. I think from from my end that's the soft summary that I wanted to bring. Uh as I said you know whoever is uh whoever is keen to I guess to open the paper you'll see much more to that. You'll see some nice charts, you know, the convex profile of trend followers, the smile shape and all that lot. Uh, it's all there. It's all there. >> No, I mean, you you mentioned it's math heavy, so it definitely rules me out. Um but um I will say uh from from from the work that um we we did uh today I will say that it's kind of a for me a little bit also of a mathematical explanation of how Katie talked about Crisis Alpha um a while back with me on on on the podcast and and what actually was the original definition of Crisis Alpha, not what we made it into in terms of narratives, but what she actually meant back then as to why maybe this strategy uh does do well when there are crises. So, um people can go back to revisit that. Maybe I don't know, maybe Katie will bring up this paper next time I speak to her uh to put it into perspective. You might be bringing this up again once you've had chance to uh to digest and and and so on and so forth. Um, but I really do appreciate uh all the hard work given the stressful uh period that you've been in um in order to come and and talk to uh to us about this um and give and give people a taste of of what these papers are about so they could go and find them and and read them themselves. So I really appreciate that uh Nick for sure. >> Of course. Um, and of course I think an even better idea would be for uh you the audience to go and leave a really nice rating and review and thanking Nick for uh the hard work that he uh put into this and to all of the co-hosts that put in uh every single week uh a lot of time in preparing for these conversations so that we can uh hopefully bring you some new and and meaningful uh angles and nuances into a topic that we've now been talking about for for many many years and that we are very passionate about. Anyways, as I mentioned also earlier, if you want the um uh the new version of the uh ultimate guide, you can find that on the uh on the top traders on blog website. Next week, I'll going to be joined by Y. Uh so, we'll see what he brings along. Uh I'm sure it'll be fun and insightful, and it's also your chance to ask him some questions. If you have some, you can email them as usual to info@ toptradersunblock.com. From Nick and me, thanks ever so much for listening. We look forward to being back with you next week. And until next time, as always, take care of yourself and take care uh of each other. >> Thanks for listening to Top Traders Unplugged. If you feel you learned something of value from today's episode, the best way to stay updated is to go on over to iTunes and subscribe to the show so that you'll be sure to get all the new episodes as they're released. We have some amazing guests lined up for you. And to ensure our show continues to grow, please leave us an honest rating and review in iTunes. It only takes a minute and it's the best way to show us you love the podcast. We'll see you next time on Top Traders Unplugged.
Nick Baltas on Trend Following in 2026: Signals, Structure & Strategy | Systematic Investor | Ep.282
Summary
Transcript
Imagine spending an hour with the world's greatest traders. Imagine learning from their experiences, their successes, and their failures. Imagine no more. Welcome to Top Traders Unplugged, the place where you can learn from the best hedge fund managers in the world, so you can take your manager due diligence or investment career to the next level. Before we begin today's conversation, remember to keep two things in mind. All the discussion we will have about investment performance is about the past and past performance does not guarantee or even infer anything about future performance. Also understand that there's a significant risk of financial loss with all investment strategies and you need to request and understand the specific risks from the investment manager about their product before you make investment decisions. Here's your host, veteran hedge fund manager Neil's Krup Larson. Welcome and welcome back to this week's edition of the systematic investor series with Nick Bolters and I Neils Castro Blaston where each week we take the pulse of the global markets through the lens of a rules-based investor. And let me also say a really warm welcome. If today is your first time you're joining us and if someone who cares about you and your portfolio recommended that you tune in to the podcast, I would like to say a big thank you for sharing this episode with your friends and colleagues. It really means a lot to us. Nick, it is wonderful to be back with you uh this week. I know it's been a hectic week. Um and it's also been a little while since you and I did one of these weekly conversations, but uh how has 2026 been treating you so far? >> So far, so so far so good, I would say. Um you know, good morning, Neils. Um I would even say still happy new year. You know, in Greece, we tend to kind of extend the wishes until until math end. Sure. >> Um it's been um >> it's been a hectic start of the year, but not necessarily for just for business reasons. Um you know, we were in Greece with family and uh I'm not sure if you saw like this kind of airspace um issue that came about um in Athens. Um long story short, there was no possibility that no flights could land or take off from the Athens airport. It was a Sunday on the 4th and we're supposed to be flying back and then flight got cancelled. you know, the entire family took us 4 days to come back at the end of the day because, you know, you had the big fat Greek return. So, all the Greeks are coming back to London uh post Christmas. So, there's no direct flight at some point or at least with availability. Uh so, that was number one. And then, you know, a bit of like um uh stomach bags and stuff, you know, have have let us down specific like know the kids. So, it's been quite of a of a rough start, I would say, but you know, we keep our smiles and uh the year has started well otherwise and you know, business goes well and um we're busy, but yeah, glad to be back. Glad to be back. It's good to have you back. When you when I know you have young kids and I when you when you mention that, it kind of reminds me of of one of uh my own travels back from from uh Christmas and New Year's um many years ago when my kids were young um and where the snow uh essentially uh while we were sitting in the plane just made it impossible to get out of Copenhagen and uh and and we end up also spending a couple of days extra to before we could get tickets. But uh there we are. there are certain things we don't uh control. >> Um, nevertheless, as we always do, it's always interesting for me to besides all the topics um that we are going to talk about and they're going to be they're super great, but um I'm always curious as been sort of what's been top of mind with uh with uh with you um since we last spoke or in the last couple of weeks. So many things are going on of course, but there anything that kind of stood out for you? am a bit uh interested if you like and and and I'm spending a bit more time these days um looking into the the way that if you like investor narrative and thematics um are impacting uh as allocation or even more so um are determining um investors uh views and and you know we discuss in this um in this podcast series for years uh momentum and trend following right and and what is momentum trend and following at the end of the day it's just perhaps um an expression of uh of slow digestion of information. >> Mhm. >> And you know we see that in the price path which allows to then document a positive or a negative move that in itself supposing that we have this kind of continued digestion of information at a slow pace continues and that's for us a tri a signal. So what happens before we kind of digest this information at a slow pace? There is clearly information that is produced and and news that is produced in the market and and and information that gets disseminated via the media and the you know the the internet. Um so that cloud of information if we end up achieving to get hold of prior to the price formation happening it's almost as if we'll get the momentum before the momentum itself plays out in the prices. So, long story short, um I'm actually quite engaged and and and focused on um um on a partnership we're doing with a kind of a a vendor if you like or or an aggregator of um of this type of data of narrative data. Um and and I'm I'm almost trying to think how this information or this type of data set can be utilized for signaling for aocation purposes, for strategy design, you know, you name it. But you know if you ask me then you know what what what keeps me excited or maybe what um keeps me thinking um in a in a productive manner that's certainly one um not a concrete answer but certainly a concrete framework or or maybe idea um that I'm I'm kind of spending my days when I'm kind of running these days or I'm doing things outside of just replying to emails or or or pitching to clients. There you are. That's the one thing. >> Yeah. No, it's kind of interesting. it it is slightly related to something that I had written down um not like something that's uh you know top of mind but still interesting. So for example, if I had if if let's just say that the narrative um had been oh Japan has increased uh its interest rates uh by a lot clearly that must mean that the yen goes up right let's just say that for an example and um I just saw a chart uh this morning um that that showed uh how much actually of an increase the the Japanese have been um seeing in the interest rates. At the same time, the yen has just done nothing in terms of strengthening. In fact, it's probably as low as it's been, more or less. So, so I I understand the interest in this thing about what if we could get hold of the momentum before it happens, so to speak. Um but I'm also reminded from last year's um price action um in a world that is getting crazier by the day how important it is not to have any kind of prediction in your in your model. Um, so I'm kind of doubling down on this um, trend following where a lot of people don't like the fact that we have no story and we have no um, view as such. Um, but just the fact that we're super disciplined at following what has happened rather than trying to get ahead of ourselves. I I probably reverse the argument a bit in the sense that I I hear you on the ex on the example about about Japan, right? And I guess after the fact u maybe the I guess the explanation there is no exposure or maybe there is no sensitivity >> to the narrative itself. And and I think that's maybe the direction I'm I'm thinking of going in in terms of there could be some story around whatever let's not name it but let's say an event the question is you know whether there is at the minimum any price reaction that we should pay attention to or not. So you know the fact that there is more intensity on a specific topic does not necessarily translate to the fact that you know price response will be significant. So I I don't have an answer, right? I'm no >> thinking out loud, right? Let's say for instance um you know equity prices are not necessarily impacted too much about inflation unless inflation is rising beyond a specific threshold. >> So it's a bit more of a conditional exposure and that exposure to the extent that we can harvest it and and and monitor it and quantify it then we might have a chance of making use of it. Um I'm not necessarily sure that there is an end to it but I'm actually kind of experimenting. The other way that I would think about it to your point and I think we've discussed it quite a few times because at the minimum I'm I'm I'm quite of a proponent of um of this idea that signal to noise quantification is important specifically for the Vshapes. So how do we how do we how do we get hold of signal to noise ratio in the market is not too easy right I mean beyond the returns scaled by volatilities which is a return over volatility or like signal to noise if you think about it in a more kind of um huristic manner we have to get hold of um um I guess non-pric information and and maybe recycling of of themes and and and narratives could allow us to get there. I don't have an answer by all means. Uh but that's a place I'm I'm quite excited about. Um just because I think we've done as much as we could with prices and volatilities and maybe there is more to be done outside of that spectrum. Maybe there isn't. But unless we experiment, you know, we we will we will not be able to to to to bring news and innovate. So that's the topic. But by all means, I think challenges are are there for us to at the minimum um um if you like entertain. So >> everything is well received. >> Speaking of narratives, you you're not I'm sure you won't believe this, but but it is it is true. Um so I was looking at my notes from last time we've recorded one of these episodes, and I don't remember exactly the date when this was, but it's been a few months ago because >> probably October maybe. >> Yeah, something like that. And so I always see what was on my radar back then. I had written down two things. One was Greenland. The other one was Trump's attack on the Fed. And I'm thinking what? I can just reuse these topics and say, "Oh, this is very much on my on my agenda." And I know you can't talk about uh politics uh of course, but um >> maybe maybe give me maybe your give your audience the signals then now like the m is m is yours. >> This is the thing. I have no idea what the narrative or what the markets would even do if we had um uh you know a massive outcome in in either of two of those two. Um but anyway and what what would actually what is interesting relating to to these two of course is that the the person at the center of attention in in both of these cases um he is coming to Switzerland to the to Davos as far as I understand. So, so, uh, there'll be lots of narratives coming out of, uh, of the of the mountainous country, uh, I'm sure in the in the next week or two. >> I'm sure. I'm sure >> people will pay attention for sure. >> Anyways, let's move on to something where we are uh, where we can talk freely, namely, namely trend following. Um, so January is actually off to a very strong start and, um, I hate to say it, so was January of 2025. Um but there should be no necessarily repeat just because we uh see both starts being very strong. Now January 2026 though nevertheless it is kind of the seventh month in a row where we're seeing very strong conditions uh for trend. We're also seeing a continuation, I think it's fair to say, of what really drove performance in the second half, namely medals uh and equities. >> Certainly setting the pace, but obviously I'd love to hear your uh kind of takeaway from um the the beginnings of 26 and and maybe you want to talk a little bit about um you know, 2025 and and then maybe later we can talk a little bit more about the QIS side of things. But just if we just stick to trend right now um what what you've experienced. >> If we stick to trend um I think you probably said it all in the sense that I think 2025 I don't have the stats but this is more of a hunch um is possibly one of the largest dispersion years we've seen. >> Mhm. >> Um without necessarily that meaning that we have seen >> um correlation breakdown between th programs. So specifically if we you know if we look into more of an outlier type of a of a scaled correlation analysis in other words to make it very very simple if a drop came in April I'm sure it came across the board maybe a different scale uh but if we assume that the scale is somehow um diluted in a correlation analysis and there are ways that we can do so throughout the year whether you're following like you know a 3 month or a six month or a 12 month trend like you know shortterm to medium to a longer term the correlations were relatively elevated anyway because to your point there were a few markets that exhibited strong trends and this is probably equities and then precious metals that's it >> y >> um and then there were pretty much the same markets that actually failed and maybe it's now um quite continued failure uh for some time you know interest rates obviously that's the one I'm talking about that's the elephant in the room >> um >> so even If the moves were captured relatively successfully if you like by a variety of trend speeds and programs the dispersion of outcomes was quite significant. I guess from that perspective, do we learn anything? You know, frankly, nothing more than just yet another year that we know we do the diagnostics and and we realize that, you know, yes, if you were too fast and possibly that Vshape did not play at your benefit or if you were too diversified, you're actually missed those few markets that did the job and the rest, you know, we're just now adding noise to your program and diversification did not play out for this year and maybe that's what we have seen the last couple of years. Are we entering a phase of should we think differently now from this point onwards? you know, the fact that we discussed that at the year end um at the year end um >> conversations, yeah, >> conversation we had with with the rest of the crew. >> Um I think the universe and the speed are probably the two most important um pillars of dispersion at least for the last year. Um, and it's fair to say that had you been more on the faster front and had you been more diversified or if you like a broader universe, probably you wouldn't be at the top ranks. But the fact that we now have a universe of very positive and very negative returns that in itself is is is an interesting observation as to how we react as product developers, how do we react as consultants, how do we react as investors, how do we react as, you know, asset owners. Let me make a pause here. >> Yeah. >> Um I can I can move on more to what we're seeing and and how we're discussing with specific investors, but maybe if you ask a question, I just put your views. I can I can I can pass it on. >> Let's do that in a second because let me it gives me a chance to do two things. One, just to give a quick uh trend update as I normally do on that, but also just to reiterate something I said to Rich last week. Um >> and by the way, I think Rich brought a really really interesting um set of topics uh last week. So for those who missed that conversation uh you know you should really go back and and and listen to that. But anyways, um, in so many ways you you talked about the dispersion, all of that, but in so many ways you could say that trend, uh, sorry, the last year was, you know, unusual, extraordinary, unpredictable, all of those acronyms that we we we we hear. But in a but in another way, the year was also very very familiar from a trend following perspective. Now I know the outcome there were differences but if you think about the fact that most managers probably made all the money in only a handful of markets and other than that we had many small losses maybe one or two bigger losses in in markets. I mean that's very classical trend following. Um so that that is to me is very comforting. um despite you know whether people ended up plus 5, plus 10 or minus 5 doesn't really matter but but the structure of how the year ended and and the return distribution all of that stuff actually look pretty pretty uh familiar to uh to um to trend followers. Anyways, we talked about uh trend uh sorry CTA coming off to a good start so far this year. So let me quickly run through the numbers. We have the Btop 50 index uh as of Wednesday evening and by the way I think yesterday was a pretty um quiet day but as of yesterday uh Wednesday evening up 374% for the Btop 50. Um so gen CTA index up uh 4 and a4% very strong trend index even more so up 4.67% and the short-term traders index up 1.75%. So all coming off uh very strong. Now, I actually completely forgot to look at the traditional markets, equity markets, and and so on and so forth. Um, but so far, just eyeballing it a little bit, obviously, we've had some positive developments in in many of the uh equity markets already this year. Um, while fixed income is probably um only slightly up, uh if anything, so we'll we'll leave it at that. Now, before we get into the topics that we're going to be talking about, uh Nick, um just uh allow me to uh mention to uh our listeners that I just published the eighth edition of uh my ultimate guide to the best investment books of all times. Um and that essentially now has more than 600 book title titles that people can uh dive into. So, it's more of a reference guide if they want some inspiration. Uh hopefully now that we put it in in one um PDF, it makes it easier for people to grab hold of. If you don't if you're not on any of my email lists, um you can go to toptradersunplugged.com/ultimate and you can get your copy and it is of course free of charge. All right. Now, I would love to um hear you talk about 2025 maybe from a QIS uh perspective. Let's break it down maybe into sort of universe versus uh speed. A little bit of continuation of the conversation we had uh with the uh the rest of the crew as you said um back in December. >> Let's do that. And and by the way, as you were speaking, I was kind of checking some of my stats for for the month to date or like year to date. Mhm. >> Um >> yeah, strong performance broadly speaking like on the on the four to five to 6% depending on the speed. Um and that's more for a 15 volt type of um of a measurement primarily driven by equities and then commod. So kind of similar things to what we saw in the kind of second half of of of the year. Let me pull out some details for for the discussion, right? Um so you asked about um I guess speed and universe. Speed is probably the most important differentiator and then universe is the second. >> If we kind of stick to the core design principles of trend following. Yes, whether it's dynamic sizing or or static is important. Yes, whether we use correlations or not is important. But long story short, assuming that there is some form of stability of volatilities and correlations, I think the speed will create more of a dispersion than the sizing. I haven't tested it but that's my hunt as it stands. Um, so we did this nice exercise which I very softly discussed in the in the in the group review. Um, that looks into fast programs. Let's say looking into like a twomonth type of a signal you know on a on a half-life basis that goes up to 12. Um, and that was kind of similar to what Quantica did, which I think also we kind of briefly touched upon back in December. Um, and then we did a similar exercise whereby we fixed the speed, uh, but then we changed the universe and we went from something very concentrated. >> Um, and I believe we use something similar to DBMF's universe just to have like, you know, an anchor which is kind of independent to our choices um, up to something that is very broad, you know, more than 80 100 markets or so. And then we tried to have some sort of mediumcaled universes maybe a bit more liquid but actually well thought through like in the sense of know if you use for instance non- US equity indices you might as well just use the EF or you can use you know UK Germany I don't know you name it like four or five major markets XUS so you can broaden up a small universe by just having maybe more tenors in specific interest rate markets or you have more in that sense. So the if you like the medium-sized universes would be reasonable rather than saying oh let me get you know 20 currencies and three equities and and we try to see here and here from for the last 25 years if you were to rank purely by annual performance however big or small the differences are which was the universe that outperformed and which was the speed that outperformed and then do some sort of a ranking which was the best speed the second best speed the third best speed and so on and so forth. there's something quite striking coming out frankly um and this is that for the last three years and that's exactly what I said in the in the December discussion if you were slow you would have out outperformed >> there is no other time in those 25 years that three years in a row being so slow would have allowed you to outperform and by the way the second best speed is just 9 months instead of like 12 months So slow has been the thing >> m >> for SVB for dollar yen for liberation day the three of these shapes you know you and I have been discussing for the last I don't know how however many months now universe wise having been small is possibly quite dominantly the outperformer for the last four years um if it's not the first to second best which is also the first time we're actually seeing it historically. There have been other times that being small was actually helpful like I think for example 2018 which I would imagine would have helped you uh towards Q4 of 2018. Um but there have been years in between whereby being very small and concentrated would have delivered the worst performance. It's only the last three four years that a concentrated universe is top ranked so stably um visav history. Right. So we have two if you like uh stylized facts that we have not seen historically and they're almost the flip side of what we saw maybe in you know in 2021 or so. >> Can I be even more explicit because you shared you shared the data with me. What's what's what really is striking to me when you talk about the portfolio size because I think and again I'm not trying to pick a fight here with my good friends in the replicator space but clearly replicators tend to use smaller universes even though that we have seen uh some some uh some new uh entrances using larger universes. But what is very interesting is as you rightly say that pretty much since 2016 small universes have been um you know the better performer. It's flipped between small and and super and super large meaning 18 markets or so. Right. Okay. But but >> dominantly it's been it's been that. >> But then what was really striking is that if you go back to the environment uh post the tech bubble. >> Yeah. Exactly. and all the way through 2009 >> exactly >> being small or trading a small universe was the worst universe every single year for something like seven years in a row. So what I'm or what I'm trying to say here is that I think we need to be very cautious of making too many conclusions based on people saying oh but look we can and here I may be rep maybe picking a little bit of a fight with replicators by by replicators coming out saying oh look at this we've now demonstrated for the last three or four or five years that we can outperform uh the CTA index. Sure, that is true and it is in line with what you're showing in terms of your analysis, right? But if we were to expand this over a 25 year period, I'm not so sure that that would have been the same conclusion based on your completely independent type of data here. >> Yeah. Yeah. I think that translates into a question which is do we learn anything uh on the back end of of this analysis? You know, in all fairness, know my mindset has always been try to be reactive, try to be diversified. >> Yeah. Did we get it wrong? I would not claim that being a wrong decision, but certainly from a performance standpoint, this is not necessarily the one that favored us, you know, made us a favor. Yes, it did in 2022. Uh at the minimum from from a reactivity standpoint, >> but um >> but that's one if I can interrupt here, that's the other thing that I find interesting about um your data when we look at the the speed, right? >> Yeah. So in 2022 you you conclude based on your analysis that actually being very were very quick uh was the best. Um and of course you can't you not you know you can't see whether it's just a you know how much of the difference is you can just see you know you can't but exactly but interestingly enough 2022 from from our side at done uh was actually one of the strongest years we had seen in in a long time. So, so that that also uh strike me as being very interesting uh as well. Um and I don't know what the short-term traders index did in 2022, but also in 2008, I do remember from our analysis that yes, the the very best single look back period was something like 56 days or whatever it was. Not a very particular long period, but every single period did well that year. I mean, you know, in a >> So, that was my point, right? That was my point. My point here is that >> so two things given we had sustained trends in 2022. >> Mhm. >> So we had the commod in the beginning rates throughout and then a dollar in the summer until November. I mean you get to it a month later or a month earlier fine. Uh it will create differences but then you know from the plus 40 to the you know plus 35 you know fine you can still rank them but in reality in terms of absolute returns that's actually quite substantial. So what this analysis is missing at this point in time is also an indication of what's maybe the spread between the top and the bottom ranked >> that I might end up kind of adding right. So to give an indication of that what is the spread in that ranking because it it might actually be like very very small for that to matter. >> Mhm. >> Um but you know setting that aside I think I I think it's a useful >> kind of um grid for us to >> to get a better sense of of the impact of those decisions. But as I keep on saying and I've discussed that with you and I've discussed that back in the days with with Morris when we talk about how we think of of str design in the QI space I think reacting to out or underperformance is is the wrong recipe right uh either way >> um I think discipline is important I think trust to the process is important um and and if there's anything to be learned frankly um it's not that a specific universe is better than another um it is more about that the Vshapes are the ones to have caused this dispersion rather than oh actually you should have been longer term because there are longerterm trends. No, that's not what's going on here. What is actually going on is something as much more subtle and it's a reversion. It's not that the that the signals and the and and and the directional moves that we have seen had like a longer duration or a shorter duration and therefore a certain speed is better than another. there are beyond trend dynamics that we should look into rather than completely amending our if you like our philosophies about strate design and and and reacting to it because you know how it's going to be right you know imagine being here in uh I don't know in whatever 2022 and and say hey I'm going to run it you know fast because you know I think that's the best and in reality you know you just make the the wrong decision at the wrong time so it's not about reading back the past but you know I think some of those fact patterns are actually quite interesting at the minimum to to to be aware of that that was the whole point of that analysis, right? >> And and I would add to that because I did my own little different type of analysis as as I knew we're going to talk a little bit about that. I was trying I was digging into some of the data I have access just to see kind of a little bit about kind of contribution by sectors because a lot of people uh in 2020 after 2022 concluded, oh yes, we should just continue to to trade mostly fixed income markets because they have been the best ones for the last 20 years and they were fantastic in 2022. We should just continue. Well, lo and behold, um pretty much since 2022, it's been kind of the worst sector to trade. Um and and so again, it just goes to this uh idea of not trying to uh uh predict too much about what the future uh will bring. Um the next thing I did was I looked at kind of all the sectors that we trade. And it's not necessarily the same um definitions that people might have. Um but it is interesting to see uh if I go back about 20 years that all sectors have actually been profitable. If you look back uh then we've been able to find opportunities in all sectors but it's clearly changing over time. Um and and um and so the the true diversification sort of the the truly diversified portfolios um I still think uh over time is a is a better is maybe a strong word to use but it's um uh to me is a more um intuitive way of doing trend following is to have the full breath of of markets in in the portfolio and not trying to >> be too concentrated. >> Well, I'm I'm in favor of of of dynamic allocation. Um, and if the opportunity set allows you to enter into a space that at least seems to be delivering some return, you might as well just go, right? So, so I see it more as an opportunity set that, you know, to the extent it's available to us, we tilt in favor or against depending on the intertemporal trend uh trend signals. So, I'm with you on this one >> perhaps with this conditioning information. >> Sure. Cool. All right. Well, we have three papers. We'll see if we get through all of them. Um, but I think we we will certainly touch on all three. Um, they're very uh different. Um, and the first paper actually they all came out I think in December uh of uh of last year. So they're relatively new. Uh the first one is probably the most accessible for most people to read. Um it's from Mikita the consultants. Um people might remember uh a paper we discussed a while back um certainly on the podcast not exactly sure who was who I spoke to about where they talked about trend following in the light of sort of crisis alpha and um they introduced this term I think they were the ones who introduced the term about first responders second responders and so on and so forth >> the RMS framework >> yes exactly and um and actually I think that was very helpful um frankly um I think it helped people to understand uh when um they should expect um positive performance during crisis periods and and so on and so forth. So I'll let you go through um this paper. We don't want to repeat uh things we've talked about too much, but still it's always useful um especially when it comes from people who probably have a lot of their or have the ear of a lot of large institutions to to uh hear about how they talk about trend and and what they're focusing on. So I'd love for you to maybe pull out some of the highlights in in in your perspective. >> Yes. Um and and as you said, that's more of an introductory paper in in in this regard. So there is um at least for for for the for the connoisseur of the space. Um it's more of a reiteration of what following is you know what's the the reason of its existence and and how we can get hold of it and so on and so forth. Um and you know speaking obviously of the RMS framework of course very well known it's pretty much kind of in line with our broad principles you know we wrote a paper back in 2019 as to how we think of defensive solutions and what are the different pillars that allow us to um to to to deliver defensive solutions from the most reactive and high basis and very expensive uh from a cost of carry standpoint to the more carry generating but not necessarily adding to the downside and obviously trend following is a kind of intermediate component uh so what how we had done independently at the time kind of aligns quite well with uh with their framework. So effectively what they came up about I think to me it's more interesting the fact that themselves put out a paper on trend uh they generally speak about themes and bigger topics and aggregation of of um of exposures and >> doing one piece on trend it's probably an indication of um I guess an increased uh interest in the space >> uh which I think was interesting for for us to at the minimum kind of bring up um they talk about what trend following is. I'm not going to go into the details. I think there's a nice section that talks about the differences in the programs which by all means I'm just going to read them out um list them effectively because we we kind of know and we have discussed them. So universe we just talk about it. >> Speed we just talk about it. >> Diversifying non signal non-trend signals. I'm going to come back to that because that's quite interesting. Stop-loss and profit uh taking triggers. um static risk allocation or dynamic risk allocation or maybe equal risk or risk budgeting between assets or asset classes that's an important one we we have also discussed it and I think I've just mentioned it earlier on on this dynamic allocation the opportunity set and then the impact of volatility targeting leverage that's another topic I think you and I discussed a few months back as to how we feel and we see the impact to leverage coming from portfolio design and um and and and Vshape dynamics and correlation shift. So they kind of mention a few of those topics as being important for um I guess for the for the design of a strategy. >> I'll probably spend a kind of tiny bit of time also kind of related to your question on on on on QIS and the CDA space from our um vantage point in 2025 which are the non-trend signals. Um I think I've been uh a proponent of of thinking about me reversion dynamics, micro indicators, K trades and so on and so forth. Um you know 2025 was a strong strong strong year for all those concepts. Um they all performed really well. They all performed in line with um at least the premise of diversifying trend following. So that did make I think a significant uh or had a significant impact in this dispersion specifically for those that implicitly use them um in the design um and I think they can have some value um not to make trend following better but allow it to be maintained in an location at times that it's actually not performing well um and and even if it's understandable under performance to your Like yes, you're holding equities. What do you think is going to happen when equities fall by 10%. Like 1 plus 1 equals two. I'm with you on this one. >> It's more the point that ultimately investors follow like a like a a an increasing concave utility function to put the economic term to it, which basically means that you know we enjoy gains less uh than the pain that we incur when a similar magnitude of a loss um is is coming. um and withholding a loss uh is tougher. So to my earlier point, should we be able to allow allocations in trend following be maintained by those non-trend components, I think we eventually make the investment process uh more complete in in the longer term and I think that's important. Um I think the last point that I would mention about the Makita paper is that they are they're actually quite open by the fact that now institutions have a variety of possibilities in terms of accessing the the style. Um of course they talk about CTAs but then they talk about mutual funds, they talk about usage, they talk about ETFs, they even talk about QIS. Um so I think to me it's a recognition that the possibilities of accessing it have widened and and and broadened. Uh but also the need for it has become more mainstream than before. Now whether that is a policy portfolio discussion or whether that is still an overlay discussion, it's um not of secondary importance but maybe of a secondary topic that is not necessarily covered here but at the minimum being there, you know, it's an indication of of of of interest and and I think I just wanted to bring it up um in our discussion. That's what it talks about. So I think people as you said uh I'm sure they'll find it very easy to go through as a as a quick overview. >> Yeah. No, absolutely. And and uh I agree. I think that um I mean you put it in this in in the sense of of diversifying uh signals to in order to maintain an allocation to trend. Um and I think also from memory at least um it's been a little while since I I read the paper in in detail but I think it's this idea also that a lot of trend following allocation kind of fails mostly because of the behavior or mostly because of the kind of reaction pattern of of investors. Um and and therefore it is definitely a uh an important uh an important point uh for sure. Now I think we are going to um move on to the little bit more hardcore uh stuff um where you need to be the translator of these uh papers. Uh it's actually two papers we're going to talk about. We'll probably come back to them given the fact that uh we we feel that they are very important. They deserve uh respect and and maybe um we haven't been able to uh devote enough time to really get into the nitty-gritty of that. I'll I'll I'll let you give your own disclaimer since you're going to be the uh the the the narrator or the translator of these um very important papers. But the first paper is by a group of distinguished people uh I'm sure um uh to be as Moscow uh who among other things uh it seems like he's linked to uh our friends at AQR uh we have someone from the Stockholm School of Economics Ricardo uh Sabatuchi Andrea Tamonei uh University of Notradam and Bian Ulf um at University of Hamburg. So this paper also came out in December of 2025. >> Um, and it's uh title is nonlinear time series momentum. That's as much as I'm going to say about it. Then I'm going to turn it over to to the expert. Um, and since I'm not a quant, I'll be allowed to ask some stupid questions along the way, I'm sure. >> Um, no, thanks. Thanks, Neils. Um I'm I'm sure this paper has appeared on people's um kind of inboxes. Um you know it came out in a sarin just before Christmas. Um of course Toby Moskovich was the you know was one of the co-authors of maybe the more um wellsighted paper when it comes to trend following in the academic literature. That's the time series momentum from 2012 >> right >> um with his colleagues at AQR at the time or in um Peterson Lassa Peterson. >> Yeah. Um so it's called kind of time you know nonlinear time series momentum. Um it's quite catchy to talk about nonlinearities in the trend following space. Um I think my honest opinion before we go into the details is that they spend time discussing a topic um but I don't think it's necessarily completely new um and this is how past information predict future information and whether that is a linear or a nonlinear transformation and I'm going to go into details. I think the novel thing that they bring about is that you know if there are those nonlinearities instead of us forcing them by some sort of a parametric approach let's say we know we utilize something that is called a sigmoid so it's a signal that goes from negative to positive but you know it flattens out in the tails their point is that we know maybe we let the data speak and maybe we use a neural network that allows us to uncover those relationships and then we can determine the positioning on the basis of um uh of those models. So let's kind of break it down um into into I guess into pieces into components. I guess let's think of um of of trading signals in the trend space. Um what are the possibilities? I would probably say that the simplest of them all is a binary signal. >> The market goes up, you buy. The market goes down, you sell. And maybe you size your exposure statically or dynamically by some level of all. Fine. Let's assume that all volatilities are the same. So you just buy a unit of oil when oil goes up and you sell a unit of S&P if S&P falls. That's it. Assuming that they have the same wall. Right? Now maybe one step of I guess of departure from that is to say well maybe if my signal is too close to zero then yes the sign is positive but in fact I don't have so much confidence on it. So I might as well just reduce my bet as a function of how big that signal is. In other words, if it's a basis point positive, then maybe I should do a basis point allocation rather than a 1% allocation. And if it's, you know, 1% of a positive return, then maybe I should do a 1% allocation. So I should scale linearly my exposure to the signal estimation. And that moves us from a binary mindset into a linear mindset. The bigger the maggot of the signal, the bigger the risk or the sizing that I should do. Um but there comes now kind of the third iteration of of of that signaling which again has been uh studied in some form or fashion. Maybe not as much in academia even though there are some nice papers that do some uh empirical analysis on that on that topic. But certainly I would want to believe in the industry. I mean certainly if I were to speak about um the work that we do um is to acknowledge the fact that yes fine you might actually use the magnitude to scale but how about a an extremely positive or an extremely negative signal. Let's say now you witness S&P up by you know three units of sharp ratio. How confident are you to size your exposure so much that these three is going to kind of repeat? >> So there comes a point whereby confidence and maybe estimation noise and maybe reversion dynamics come into play and then you kind of say like you know what there is a point whereby I just want to flatten out my exposure. So I get it, the higher the better, but you know, probably there should be some sort of a concave form on the positivity side and more of a kind of a convex profile on the negative side that flattens out the exposure. So I still maintain my linearity around zero. So positives are positive, negative, and negatives. But then in the extremes, I should just flatten it out. And I call it a sigmoid. Um, you know, they call it the vapor sshaped. Um I'm using the kind of the Greek if you like translation of what um of what an S is called like the sigma. Um so that's if you like the third iteration. So again to repeat binary long or short linear just use the magnitude sigmoid maintain the magnitude but then bring some nonlinearity. Um and if we stick to those three before we go to what the paper really really focuses on, there is some form of um of a benefit performance-wise. If you look into empirical analysis, going into that more like of a sigmoid type of mindset or going into this kind of linear space, um maybe the only places that you'll see some loss of performance is when a tiny bit of a trend is enough for you to get full into the position with a binary signal and then benefit from an early start. >> Mhm. >> But imagine a signal that goes bit positive, bit negative, a bit positive, bit negative. with two full positive and two full negative exposure you'll end up acrewing a lot of costs. So net of costs it can be a debate how the binary signal can actually help. So this transition is always a bit helpful. Um so there is empirical evidence and they also showcase it that you know this nonlinearity even if we force it to be like a sigmoid and there are variety of ways that we can produce that mathematically and and parametrically um you get good performance or better performance. Their point is why would you have to force the shape of that sigmoid? As in everything I've discussed so far is not necessarily driven by the data. It's almost as if we're forcing it. So the sign of past return is our determination. The linear is our determination. The smoothening of the tails is our determination. And the signal we put in place is our design. And yes, maybe after the fact we look into the empirical stats and say hey you're actually doing better that in itself almost says that yes probably the data generating process is such whereby you have this continuation of return but in the tails it ends up reverting. So they say why don't we let the data speak. >> Mhm. So instead of us scaling up and then smoothing it in the tails or scaling down on the negatives and then smoothing it out in the tails, let different markets you know be it equities or rates or commodities and so on and so forth force this nonlinearity while still preserving and that's important the fact that the signal is positive I buy the signal is negative I sell. And that's important why because the minute we depart from it it is no longer just a trend strategy. Imagine me being here and telling you, "Hey, your signal is positive. You're going to short." Well, this is no longer a trend strategy. It's something of a mixture. >> So, they use a neural network. I don't want to bother you too much with the details, but they use a neural network to effectively um um if you like extract the way that past return translates into future return, right? What's the what's the transfer function? How do I um document a signal? And how do I determine my position? And should my position be linearly related to my signal? Should it be fading out if it's too extreme? Should it maybe just fall? Let's say to my earlier point, if my sharp ratio is three, should I possibly just go to zero? I think crossing zero and going to negative. I see the value of that because if the data tells you you need to actually go short then it basically tells you that there are reversion dynamics. Now whether we bring the reversion dynamics into a trend system or we use them as a separate engine to my earlier point that's maybe cleaner from an estallocation and from an from from a performance attribution standpoint but they do show I guess to to go to go probably closer to to to the end of um of my overview that um empirically the data spits out transfer functions that have the nonlinearity that we're discussing which is the smoothening in the tails but even you can end up having reversion dynamics as well in the very extremes. Um now I said that those pictures of fitting data at the end of the day just like know if you put past returns future returns and do like a nonp parametric analysis you'll see it. So there are some papers a couple of years back I believe some of them we have discussed here but maybe my memory is kind of failing me now that show this relationship but what they say is that they are now actively monitoring it and fitting it to to get to that point. So to conclude basically the paper says look there are nonlinearities those nonlinearities exist in the tails um this is something that we should at the minimum be aware of whether we design that parametrically or nonp parametrically it's a good question whether the neural network is bringing value maybe it brings value um you know that's why one of the findings that you know the the neural network agrees with the theory um and the fact that you know in the tails possibly the estimation error is creator and so on and so forth and some reversion dynamics are there Um I would probably say it is an important discussion to be had as to whether this incremental complexity will bring significant value versus having a sigmoid that in itself kind of be in itself behaves um in line with with with expectations. So I leave it uh in in this regard but by all means you know very good work and and and I'm sure people will will spend time looking at it and maybe trying it. >> Okay. Well, the last paper um is from a uh a fellow Dane actually. Um so uh um so we're going to be talking about a paper from a gentleman called Christian who works at the uh one of the largest pension funds in Denmark called ATP and he also posted a new paper in December called on the autonom autonomy of trend. So, um I know we're going to have a hard stop in our conversation today. So, I'm going to let you talk as much or as little uh about it. I think it is uh important because it touches on some of the key things that people actually uh consider uh when they incorporate trend following into their portfolio. Uh and that is kind of the again the defensive uh nature and mechanisms. So, u I'll let you uh I'll let you talk about it uh Nick and we'll see how we go. Yeah. Yeah. Yeah. So that's um that's a nice one by by Christian. So I I I I also happen to know Christian. Um and um it's a nice piece of work. It's actually quite mathematical. Uh it's more theoretical. Um but I think what is what is important at times with those theory papers is to kind of look through um how beautiful sometimes math can be and kind of identifying what we experience in reality and kind of giving a bit of a fundamental base if you like. Um so his point is more about okay what is actually driving trend following and and how do they behave across different market environments and what is the underlying process of the data that gives rise to trend following um and and and somebody could argue yet that you have autocorrelation returns you know positive is leading to positive negative negative there are some good papers from back in the days that showcase that even if returns are independent if volatility in itself is serally dependent you can still benefit from a trend following strategy uh I'm not going to go into the details of this one but you So that was the first that came to mind when I saw when I saw the paper. Um they you know he he he kind of built a model you know he says that no log returns are following a ocean process. Um you know we we we design a trend signal which is just the summation of past the returns. I'm not going to bother you too much. You know they're using a binary signal u to go to to back to to our previous conversation. But then what what he what what they try or what he tries to to come about is um a mathematical explanation of what is the expected payoff of a trend following strategy unconditionally and then importantly conditionally on what has happened in the in the recent past which is more of a of a trend following behavior in itself um specifically condition upon um some cumulative return of a specific horizon. And I think the direction of travel here is to say well if I'm observing a significant draw down what is my expected payoff from being a trend follower. Um and there are two kind of components that come out of this analysis which is one being the directional component. By all means we know what that means. You know if you get on average the direction right you will do well. So if you're holding equities more often than not with an equity risk premium that is positive over the longer term you'll probably do well. And that was I think the challenge we had with equities and bonds you know three four five years ago. I'm sure you remember the days that everyone said, "Hey, you're buying equities, you're buying bonds, you're just benefiting from carry and equities premium." No longer the case. I think that that that doubt is is is out of the window for now. Uh but then the second element which is also more important for for us is the timing element. So how quickly can you turn into a situation whereby your um price process um is aligned in terms of direction with your signaling. Um and I think one of the more interesting things that you know his analysis uh showcases is that you know beyond the dependence that we have and we know that you know the getting the signal direction right is is is the recipe for success and so on so forth. Um there's also this part of the analys that says even if returns of the process you're trying to allocate to on the base of trend following is um is what we call iid. So in other words independent identically distributed. So every single day the equity returns are drawn from the same distribution but today's value has no dependence to yesterday's value and so on and so forth. So whatever happens every single day is independent. Conditional upon observing a drawdown which can be a consequence of independent draws like you know I can do heads or tails 10 times and I can get heads 10 times. They are still independent right. But it can happen. >> Sure. Um they show that you know um beyond specific thresholds mathematically defined a trend follower can still um deliver positive return and specifically in the draw down specifically as a crisis alpha uh which goes kind of hand inhand >> um with um with some of the empirical observations. So I kind of I know I went a bit quick um in the interest of time but two things to flag. It is a mathematically heavy paper which is actually quite nice. Um but I think what's more important is to associate now the findings that um if you like formula and and variables um if you like contain to to the experience we observing and and and I think going past the fact that you know serial correlation is important and the fact that even a an IID process can allow you still to be profitable with a trend following strategy is an important finding. Again, it reminds me of that work that says conditional dependence in the volatility space is enough for you to benefit from a trend following strategy or in other words to predict the directionality. Um so I don't know if you have any reflections upon that. I think from from my end that's the soft summary that I wanted to bring. Uh as I said you know whoever is uh whoever is keen to I guess to open the paper you'll see much more to that. You'll see some nice charts, you know, the convex profile of trend followers, the smile shape and all that lot. Uh, it's all there. It's all there. >> No, I mean, you you mentioned it's math heavy, so it definitely rules me out. Um but um I will say uh from from from the work that um we we did uh today I will say that it's kind of a for me a little bit also of a mathematical explanation of how Katie talked about Crisis Alpha um a while back with me on on on the podcast and and what actually was the original definition of Crisis Alpha, not what we made it into in terms of narratives, but what she actually meant back then as to why maybe this strategy uh does do well when there are crises. So, um people can go back to revisit that. Maybe I don't know, maybe Katie will bring up this paper next time I speak to her uh to put it into perspective. You might be bringing this up again once you've had chance to uh to digest and and and so on and so forth. Um, but I really do appreciate uh all the hard work given the stressful uh period that you've been in um in order to come and and talk to uh to us about this um and give and give people a taste of of what these papers are about so they could go and find them and and read them themselves. So I really appreciate that uh Nick for sure. >> Of course. Um, and of course I think an even better idea would be for uh you the audience to go and leave a really nice rating and review and thanking Nick for uh the hard work that he uh put into this and to all of the co-hosts that put in uh every single week uh a lot of time in preparing for these conversations so that we can uh hopefully bring you some new and and meaningful uh angles and nuances into a topic that we've now been talking about for for many many years and that we are very passionate about. Anyways, as I mentioned also earlier, if you want the um uh the new version of the uh ultimate guide, you can find that on the uh on the top traders on blog website. Next week, I'll going to be joined by Y. Uh so, we'll see what he brings along. Uh I'm sure it'll be fun and insightful, and it's also your chance to ask him some questions. If you have some, you can email them as usual to info@ toptradersunblock.com. From Nick and me, thanks ever so much for listening. We look forward to being back with you next week. And until next time, as always, take care of yourself and take care uh of each other. >> Thanks for listening to Top Traders Unplugged. If you feel you learned something of value from today's episode, the best way to stay updated is to go on over to iTunes and subscribe to the show so that you'll be sure to get all the new episodes as they're released. We have some amazing guests lined up for you. And to ensure our show continues to grow, please leave us an honest rating and review in iTunes. It only takes a minute and it's the best way to show us you love the podcast. We'll see you next time on Top Traders Unplugged.